一种带松弛因子的统计粗糙集挖掘算法

Mining algorithm based on statistical rough sets with relaxation

针对实际应用领域数据挖掘的数据特征提出了基于底层数据信息归纳的带松弛因子的统计粗糙集数据挖掘算法DMSRIIR。该算法引进了保持原始底层数据信息的机制以及求近似约简的松弛因子 ,可以有效地避免生成规则的偶然性以及解决部分分类的问题。DMSRIIR还提供了应用粗糙集于实际数据挖掘的系统方法。实验表明了算法的可行性与有效性。

This paper presented a data mining algorithm DMSRIIR based on statistical rough sets model and the information induction of the primitive data. A mechanism to keep the low level data information and a relaxation factor to compute the approximate reduct were introduced. DMSRIIR can avoid generating the casual rules, and can solve the partial classification problem. DMSRIIR also provides a systematic method to use rough set mining various data sets. Experiments with UCI's databank show its validity and efficiency.